Seiberg-Witten Theory and AGT Relation Tohru Eguchi We consider N = - - PowerPoint PPT Presentation

seiberg witten theory and agt relation
SMART_READER_LITE
LIVE PREVIEW

Seiberg-Witten Theory and AGT Relation Tohru Eguchi We consider N = - - PowerPoint PPT Presentation

Seiberg-Witten Theory and AGT Relation Tohru Eguchi We consider N = 2 supersymmetric gauge theories in 4- dimensions and study the case when the theory possesses the conformal invariance. Simplest example of a conformal invariant theory:


slide-1
SLIDE 1

”Seiberg-Witten Theory and AGT Relation”

Tohru Eguchi We consider N = 2 supersymmetric gauge theories in 4- dimensions and study the case when the theory possesses the conformal invariance. Simplest example of a conformal invariant theory:

SU(2) gauge theory with Nf = 4 hypermultiplets

We may consider its generalizations A chain of SU(2) gauge theories with bifundamen- tals and fundamental at the ends: quiver gauge theories

slide-2
SLIDE 2

As is well-known, such quiver theories are obtained using the brane construction as shown in the figure: One has n + 1 NS5 branes and a pair of D4 branes are sus-

slide-3
SLIDE 3

pended between neighbouring NS5 branes giving rise to

SU(2)1 × SU(2)2 · · · × SU(2)n gauge symmetry. Two D4

branes at extreme left and right extend to x6 = ±∞ repre- senting fundamental hypermultiplets. In such a configuration each SUi(2) theory couples to Nf = 4 hypermultiplets and is conformally invariant. Thus there exists a set of marginal parameters in the theory

{τi = θi π + 8iπ g2

i

, i = 1, , n}

Uplifting this brane configuration to 11 dimensions

= ⇒ M theory picture with an M5 brane wrapping a Riemann

slide-4
SLIDE 4

surface (cylinder) with punctures. Thus, conformal N = 2 theories

≈ an M5 brane wrapping a Riemann surface C with

a number of punctures. Number of parameters of Riemann surface Cg,n of genus g

slide-5
SLIDE 5

with n punctures:

3g − 3 + n

This agrees with the number of gauge theory parameters {τi}. Hence one expects Gaiotto S-duality group of quiver gauge theory = mapping class group of Riemann surface Cg,n

slide-6
SLIDE 6

Remarkable observation Alday,Gaiotto,Tachikawa AGT relation

⟨ ∏ Vmi(τi)⟩ = ∫ [da] |ZNek(τ; a; m, ϵi)|2

Liouville Nekrasov partition function correlation function

  • f SU(2) gauge theory in Ω background

Liouville momentum

    

external line: mi

, ∆i = mi (Q − mi )

interbal line:

α = Q 2 + a

slide-7
SLIDE 7

Background charges

Q = b + 1 b, c = 1 + 6Q2 ϵ1 = b, ϵ2 = b

! "# "$ "% "&

slide-8
SLIDE 8

Nekrasov formula Sum over Yang tableau; Y = (λ1 ≥ λ2 ≥ · · · )

ZNek = ∑

(Y1,Y2)

q|⃗

Y |Zvector(⃗

a, ⃗ Y )Zantifund(⃗ a, ⃗ Y , m1) ×Zantifund(⃗ a, ⃗ Y , m2)Zfund(⃗ a, ⃗ Y , −m3)Zfund(⃗ a, ⃗ Y , −m4)

slide-9
SLIDE 9

Here

Zvector(⃗ a, ⃗ Y ) = ∏

i,j=1,2

s∈Yi

( aij − ϵ1LYj(s) + ϵ2(AYi(s) + 1) )−1 × ∏

t∈Yj

( aji + ϵ1LYj(t) − ϵ2(AYi(t) + 1) + ϵ+ )−1 Zfund(⃗ a, ⃗ Y , µ) = ∏

i=1,2

s∈Yi

(ai + ϵ1(ℓ − 1) + ϵ2(m − 1) − µ + ϵ+) Zantifund(⃗ a, ⃗ Y , µ) = ∏

i=1,2

s∈Yi

(ai + ϵ1(ℓ − 1) + ϵ2(m − 1) + µ) ϵ+ = ϵ1 + ϵ2, aij = ai − aj. LY (s) and AY (s) are leg and

arm length of the site s.

slide-10
SLIDE 10

Nekrasov formula is obtained by summing over contributions from fixed points in the ADHM formula under gauge and Lorenz transformation (SO(4) = SU(2)L × SU(2)R ∈ (ϵ1, ϵ2)). First exact relationship between 4-dim CFT and 2-dim CFT. Higher rank generalization: Toda theories

slide-11
SLIDE 11

♠ Attempts at direct proof

Fateev-Litvinov Detailed study of the algebraic structure of conformal block in Liouville theory , i.e. the recursion relation by Al.B.Zamolodchikov.

⟨Vα⟩ ≈ F∆

α (q),

F∆

α (q) =

∑ qmn Rn,m ∆ − ∆m,n F∆m,−n

α

(q)

and comparison with the sum over Yang tableaus of gauge theory side. Conformal block of 1-point function in Liouville theory on a torus = N = 4 gauge theory perturbed by the mass of the adjoint hypermultiplet (N = 2∗ theory)

slide-12
SLIDE 12

♠ Exact Integration

Consider Liouville correlation function in free field represen- tation

⟨ ∏

a

eimaϕ(qa)

N

i=1

∫ ebϕ(zi) dzi⟩

screening ops.

= ∏

a<b

(qa − qb)2mamb ∫ ∏

i,a

dzi (zi − qa)−2ibma ∏

i<j

(zi − zj)−2b2, ∑

i

ima + Nb = Q

Dotesnko-Fatteev integral

slide-13
SLIDE 13

This is an integration of Selberg type.

IN(a, c, β) = ∫

N

i=1

dxi ∏

i<j

(xi − xj)2β

N

i=1

xa

i (1 − xi)c

=

N−1

j=0

Γ(a + 1 + jβ)Γ(c + 1 + jβ)Γ(1 + (j + 1)β) Γ(a + c + 2 + (N + j − 1)β)Γ(1 + β)

Attempts at exact evaluation and comparison with conformal blocks. Morozov-Kironov-Shakirov, Itoyama-Oota· · ·

slide-14
SLIDE 14

♠ Monodromy transformations

SW curve Σ :

x2 = ϕ2(z), ϕ2 has double poles ≈ m2

i

(z − zi)2 1 2πi

  • Ai

xdz = ai, 1 2πi

  • Bi

xdz = ai

D, ai D =

1 4πi ∂F ∂ai

In the semi-classical limit → 0,

ϵ1,2 << ai, mi Z ≈ exp ( −F (ai) 2 )

slide-15
SLIDE 15

Liouville stress tensor T (z)

⟨T (z)Vm1(z1) · · · Vmn(zn)⟩ ≈ −1 2 ϕ2(z)⟨Vm1(z1) · · · Vmn(zn)⟩ ⇓ −m2

i

2

(z − zi)2 ≈ ∆i (z − zi)2

Degenerate field Consider a field Φ2,1(z) = e

−b 2 ϕ(z) which possesses a de-

generacy at level 2

∂2

z Φ2,1(z) = −b2 : T (z)Φ2,1(z) :

slide-16
SLIDE 16

Correlation function with an extra insertion of Φ2,1

Z(ai; z) = ⟨Φ2,1(z)Vm1(z1) · · · Vmn(zn)⟩

In the semi-classical limit

Z(ai; z) ≈ exp ( −F (ai) 2 + bW (ai; z)

  • + · · ·

)

One finds

(∂W )2 = ϕ2(z) = x(z)2

Hence

W±(z) = ± ∫ z

z∗ xdz

slide-17
SLIDE 17

shift around A, B cycles gives

Z(ai; z + Aj) = exp(2πib

  • aj)Z(ai; z)

Z(ai; z + Bj) = exp(2πib

  • aj

D)Z(ai; z)

Similarly we may consider the process

  • 1. Insert identity operator inside the Liouville correlator
  • 2. Φ2,1 ⊗ Φ2,1 ≈ 1
  • 3. Transport one of Φ2,1’s around A, B cycle
  • 4. Pair annihilate two Φ’s into identity
slide-18
SLIDE 18

= ⇒ L(γ)Fα = cos(πb(2α − Q) cos(πbQ) Fα

These processes give monodromy factors corresponding to the action of Wilson loop, ’t Hooft loop and surface operators. Alday-Gaiotto-Gukov-Tachikawa-Verlinde Drukker-Gomis-Okuda-Teschner

slide-19
SLIDE 19

♠ Matrix Model

Dotsenko-Fatteev integral when b = i suggests a matrix model interpretation with an action

S = ∑

a

ma log(M − qa)

and {zi} are identified as matrix eigenvalues. Dijkgraaf-Vafa We find that this model in fact reproduces Seiberg-Witten the-

  • ry (also for the asymptotically free cases Nf = 2, 3). But it

still has mysterious features. T.E.-Maruyoshi

slide-20
SLIDE 20

Let us consider the simple case of 4 hypermultiplets with masses

m±, ˜ m±. Define m0 = 1 2(m+ − m−), m1 = 1 2( ˜ m+ − ˜ m−) m2 = 1 2(m+ + m−), m3 = 1 2( ˜ m+ + ˜ m−)

Condition:

i

mi = 2gsN

slide-21
SLIDE 21

M theory curve is given by

CM : (v − m+)(v − m−)z2 +c1(v2 + Mv − U)z + c1(v − ˜ m+)(v − ˜ m−) = 0

For convenience, set c1 = −(1 + q), c2 = q. By shifting v to eliminate the linear term and setting v = xz

CM : x2 = ( m2z2 + (1 + q)M

2 z + m3q

z(z − 1)(z − q) )2 +(m2

0 − m2 2)z2 − (1 + q)Uz + (m2 1 − m2 3)q

z2(z − 1)(z − q)

slide-22
SLIDE 22

Seiberg-Witten differential behaves at a pole as

λSW = xdz 2πi ≈ m∗ z − z∗

Mass appears at residues. Pole at z = 0, z = ∞; residue ±m1, ±m0. Require pole at z = 1 with residue ±m2 and z = q with residue ±m3 =

⇒ M = −2q 1 + q(m2 + m3) ♣ UV and IR gauge coupling constant

slide-23
SLIDE 23

Standard SW curve of Nf = 4 in massless case

CSW : y2 = 4x3 − g2ux2 − g3u3

Here

g2(ω1, q) = ( π ω1 )4 1 24 ( ϑ3(q)8 + ϑ2(q)8 + ϑ4(q)8) , g3(ω1, q) = ( π ω1 )6 1 432 ( ϑ4(q)4 − ϑ2(q)4) × ( 2ϑ3(q)8 + ϑ4(q)4ϑ2(q)4)

On the other hand M theory curve in the masssless limit is

slide-24
SLIDE 24

given by

CM : x2 = − (1 + q)U z(z − 1)(z − q′)

Here U is related to u = trϕ2 as

U = Au

and we have used q′ in order to distinguish it from q of CSW . By comparing the periods we find

q′ = ϑ2(q)4 ϑ3(q)4, A = 1 ϑ2(q)4 + ϑ3(q)4

slide-25
SLIDE 25

We regard q in SW curve as the gauge coupling in the infra- red regime q = qIR and q′ in M theory curve as the ultra- violet gauge coupling constant q′ = qUV . Relation

qUV = ϑ2(qIR)4 ϑ3(qIR)4

has been obtained by various authors. Grimm et al, Marshakov et al

♠ Matrix model and modular invariance

Equation of motion

∑ mi λI − qi + 2gs ∑

I̸=J

1 λI − λJ = 0

slide-26
SLIDE 26

We have q1 = 0, q2 = 1, q3 = qUV . Eigenvalue distribution is as given in the figure.

slide-27
SLIDE 27

Resolvent of the theory is defined by

Rm(z) = gsT r 1 z − M

and satisfies the loop equation

⟨Rm(z)⟩2 = −⟨Rm(z)⟩W ′(z) + f(z) 4 f(z) = 4gsT r ⟨ W ′(z) − W ′(M) z − M ⟩ =

3

i=1

ci z − qi

Matrix model curve (spectral curve) is defined by the dis-

slide-28
SLIDE 28

criminant of the loop eq.

Cspec.curve : x2 = W ′(z)2 + f(z) = (m1 z + m2 z − 1 + m3 z − q )2 + (m2

0 − ∑ i m2 i )z + qc1

z(z − 1)(z − q)

  • Eq. of motion =

⇒ ∑

i

ci = 0

Residue at ∞ being ±m0 =

⇒ c2 + qc3 = m2

0 − (

∑ mi)2

Then

qc1 = (1 + q)m2

1 + (1 − q)m2 3 + 2qm1m2 − 2qm2m3

+2m1m3 − (1 + q)U = ⇒ CW = Cspec.curve

slide-29
SLIDE 29
  • Modular invariance

Consider the massless limit of spectral curve

x2 = − (1 + q)U z(z − 1)(z − q) = −

u θ4

3

z(z − 1)(z − q)

This is invariant under

I : (z, x) → (1 − z, x), q → 1 − q, u → −u, S II : (z, x) → (1 z, −z2x), q → 1 q, u → u, ST S

Recall q = θ4

2

θ4

3

.

slide-30
SLIDE 30

Consider massive case. Under the S- and STS-transformations mass parameters are transformed into each other

I : (0, 1, q, ∞) → (1, 0, 1 − q, ∞), m1 ↔ m2 II : (0, 1, q, ∞) → (∞, 1, 1 q, 0), m0 ↔ m1

Under these transformations, the spectral curve should be in-

  • variant. By imposing the conditions

x2(z; m0, m1, m2, m3; q) = x2(1 − z; m0, m2, m1, m3; 1 − q) x2(z; m0, m1, m2, m3; q) = 1 z4x2(1 z; m1, m0, m2, m3 : 1 q)

slide-31
SLIDE 31
  • ne can completely fix the mass dependence of the param-

eter U. Solution to the above conditions is given by

(1 + q)U = u ϑ4

3

− q(m2 + m3)2 + 1 + q 3  

3

i=0

m2

i

 

  • Asymptotically free theory with Nf = 3

precise relationship between u and T rϕ2

slide-32
SLIDE 32

u = ⟨T rϕ2⟩ − 1 6(ϑ4

4 + ϑ4 3) 3

i=0

m2

i .

Recall

m± = m2 ± m0, ˜ m± = m3 ± m1,

We take the limit

˜ m− → ∞, q → 0,

with

˜ m−q = Λ3

fixed

slide-33
SLIDE 33

Matrix action reduces to

W (M) = ˜ m+ log M − Λ3 2M + m2 log(M − 1).

the spectral curve for Nf = 3 theory becomes

x2 = Λ2

3

4z4 − ˜ m+Λ3 z3(z − 1) − u − (m2 + 1

2 ˜

m+)Λ3 z2(z − 1) + m2 z(z − 1) + m2

2

z(z − 1)2 − m2Λ3 z2(z − 1).

Predicts the same free energy and discriminant as that of the

slide-34
SLIDE 34

standard SW curve

x2 = z2(z − u) − 1 4Λ2

3(z − u)2

−1 4(m2

+ + m2 − + ˜

m2

+)Λ2 3(z − u) + m+m− ˜

m+Λ3z −1 4(m2

+m2 − + m2 − ˜

m2

+ + ˜

m2

+m2 +)Λ2 3

  • Asymptotically free theory with Nf = 2

Matrix action:

W (M) = ˜ m+ log M − Λ2 2M − Λ2M 2

slide-35
SLIDE 35

Spectral curve:

x2 = Λ2

2

4z4 + ˜ m+Λ2 z3 + u z2 + m+Λ2 z + Λ2

2

4

Computation of free energy

gsN1 = 1 4πi

  • x(u)dz

(N1 denotes the filling fraction of the first cut ”1”). Derivative of free energy in Λ2 is given by

Λ2 ∂F ∂Λ2 = −Λ2 gs 2 ⟨ ∑

I

( 1 λI + λI ) ⟩ = 2u + Λ2

2 − m2 + − ˜

m2

+

slide-36
SLIDE 36

On the other hand

4πi∂(gsN1) ∂u =

  • C1

dz √ P4(z)

where

P4(z) = z4 + 4 ˜ m+ Λ2 z3 + 4u Λ2

2

z2 + 4m+ Λ2 z + 1

This is a complete elliptic integral and we can expand u in terms of a = 2gsN1 (we put m+ = ˜

m+ ≡ m for simplicity) u = a2 + m2 2a2Λ2

2 + a4 − 6m2a2 + 5m4

32a6 Λ4

2 + · · ·

slide-37
SLIDE 37

Then by integrating over Λ2 we finally obtain

4Fm = 2(a2 − m2) log Λ2 + a2 + m2 2a2 Λ2 + a4 − 6a2m2 + 5m4 64a6 Λ4 + · · ·

This gives the same free energy as the standard SW curve

x2 = (z2 − 1 4Λ4

2)(z − u) + m+ ˜

m+Λ2

2z − 1

4(m2

+ + ˜

m2

+)Λ4 2

slide-38
SLIDE 38

♠ Discussions

  • 1. We want a much wider class of correspondences:

Liouville, Toda

= ⇒ WZW, cosets, parafermions etc. N = 2 Yang-Mills on R4 = ⇒ on ALE spaces, rational surfaces?

  • 2. Want five-dimensional version of AGT. It is known that 5-

dimensional Nekrasov formula counts the number of holo- morphic curves in non-comapct CY manifolds (geometric engineering). 5-dim. AGT =

⇒ CFT acting on the space of

Gromov-Witten invariants.